Breaking encryption with a quantum computer just got 10 times easier

The amount of quantum computing power needed to crack a common data encryption technique has been reduced tenfold. This makes the encryption method even more vulnerable to quantum computers, which may be able to reach the reduced size within the decade.

The RSA algorithm is one of the most widely used encryption algorithms, used for things like online banking and secure communication. It is based on the mathematical difficulty of finding which two prime numbers were multiplied together to create a very large number. Since the 1990s researchers have known that this difficulty can be side-stepped by using a quantum computer, but the possibility was considered theoretical because the size needed for such a quantum computer was much larger than could be built.

This has slowly started to change as researchers built larger quantum computers and the estimated size needed has come down. In 2019, Craig Gidney at Google Quantum AI co-authored a paper that reduced these requirements from 170 million to 20 million quantum bits, or qubits. And in 2025, Gidney devised a way to slash that number to less than a million qubits. Now, Paul Webster at Iceberg Quantum in Australia and his colleagues have managed to decrease the number even further to about 100,000 qubits.

The researchers’ study builds on Gidney’s work in terms of algorithmic improvements, but they assume that a different scheme is used for connecting and arranging qubits called qLDPC code. In past schemes, qubits can only interact with their nearest neighbours but qLDPC code means they can interact with qubits that are further away. This approach increases the connectivity and effectively increases the density of information within the quantum computer.

Given this connectivity, the team estimated that for 98,000 superconducting qubits, like those currently made by IBM and Google, it would take about a month of computing time to break a common form of RSA encryption. Accomplishing the same in a day would require 471,000 qubits.

Several quantum computing firms are aiming to build quantum computers with hundreds of thousands of qubits within the decade and the new estimate is largely agnostic to what they would be made from, just relying on their error rates and quantum computer’s speed. Putting aside the practicality of running a computation for a month, could Iceberg Quantum’s scheme actually be implemented in practice? Anyone in charge of a quantum computer that could do so would have access to many emails, bank accounts or even confidential government files protected with RSA encryption.

“These stricter demands make the hardware harder to make, and making the hardware is already the hardest part,” says Gidney. Similarly, Scott Aaronson at the University of Texas at Austin wrote on his blog that his main reservation with the new estimate is the difficulties in practically engineering the necessary connections between distant qubits.

IBM’s researchers have championed qLDPC codes in recent years and they have made the firm’s quantum computing hardware more amenable to them, but how successful this approach can be remains unclear. A spokesperson for IBM said in a statement that qLDPC codes will be a “cornerstone” of its quantum computers but did not comment on whether the new scheme could be realised.

Connections between distant qubits are much easier to implement when they are made from extremely cold atoms or ions, two quantum computing approaches that have gained prominence in recent years. But these quantum computers also work more slowly, which, according to the new study, may put their numbers back in the millions when it comes to breaking RSA encryption.

“I think it’s important to never be conservative with the timelines for things like this happening,” says Lawrence Cohen, also at Iceberg Quantum. “Somebody breaking RSA would have big consequences, and it’s always much, much better to err on the side of this could very much happen sooner rather than later.”

He says that breaking RSA encryption is a well-studied problem and therefore a great benchmark for anyone looking to build a powerful quantum computer, but his team’s approach could also be used to run better and more useful simulations of quantum materials and quantum chemistry.